
import pandas as pd
import numpy as np

# Modeling
import statsmodels.api as sm
from pygam import GAM, ExpectileGAM, s, l, f
from sklearn.preprocessing import LabelEncoder

# Visualization
import plotly.express as px
from plotnine import *
import textwrap

import warnings
warnings.filterwarnings('ignore')


data = pd.read_csv(r"C:\Users\arham\Desktop\lab_87_price_optimization_python\data\price_optimization.csv")


data


data.describe(include='all').T


# 2.0 DATA PREPARATION
# --------------------

data['is_event'] = (data['event'] != "No Promo").astype(int).astype(str)

data['revenue'] = data['price'] * data['quantity_sold']

data


# TRENDS: PRICE vs QUANTITY SOLD
fig = px.scatter(
    data,
    x='price',
    y='quantity_sold',
    color='product',
    opacity=0.6, 
    trendline="lowess",  
    trendline_color_override="blue",
    template="none",
    title='Product Sales: Price vs Quantity Analysis',
    width=800, 
    height=600,
).update_traces(
    marker=dict(size=7),
    hoverlabel=dict(font=dict(size=10)),
).update_layout(
    legend_title_text='Product',
    title_font=dict(size=16),
    legend_font=dict(size=10),
).update_xaxes(
    title_text='Price',
    title_font=dict(size=10), 
    tickfont=dict(size=10),
).update_yaxes(
    title_text='Quantity Sold',
    title_font=dict(size=10), 
    tickfont=dict(size=10),
)

fig


# 4.0 EVENT ANALYSIS (OUTLIERS)
# ----------------------------

# SCATTER PLOT
fig = px.scatter(
    data,
    x='price',
    y='quantity_sold',
    color='event',
    facet_col='product',
    facet_col_wrap=2,
    facet_col_spacing=0.1,
    facet_row_spacing=0.1,
    opacity=0.6, 
    trendline="lowess",  
    trendline_color_override="blue",
    template="none",
    title='Product Sales: Event Analysis',
    width=800, 
    height=700,
).update_traces(
    marker=dict(size=7),
    hoverlabel=dict(font=dict(size=10)),
).update_layout(
    legend_title_text='Product',
    title_font=dict(size=16),
    legend_font=dict(size=10),
).update_xaxes(
    title_text='Price',
    title_font=dict(size=10), 
    tickfont=dict(size=10),
    matches=None,
).update_yaxes(
    title_text='Quantity Sold',
    title_font=dict(size=10), 
    tickfont=dict(size=10),
    matches=None,
)

for annotation in fig['layout']['annotations']: 
    annotation['font'] = dict(size=10) 
    
fig.for_each_xaxis(lambda axis: axis.update(showticklabels=True))
fig.for_each_yaxis(lambda axis: axis.update(showticklabels=True))

fig


# 6.0 GENERAL PRICE OPTIMIZATION: GAMs
# ----------------------------------
# PROBLEM: Prices are non-linear
# SOLUTION: Use GAMs (Generalized Additive Models)
# NOTES:
# - GAMs are like Linear Regression, but allow for non-linear relationships
# - NOT as useful for incorporating events (see next section)

# Keep the entire data as data_filtered
data_filtered = data.query('event == "No Promo"')

# Create a list of unique products
unique_products = data_filtered['product'].unique()

# Create an empty dataframe to store the concatenated results
all_gam_results = pd.DataFrame()

# Loop through each product
for product in unique_products:
    # Filter data for the current product
    product_data = data_filtered[data_filtered['product'] == product]
    
    X = product_data[['price']]
    y = product_data['quantity_sold']
    
    quantiles = [0.025, 0.5, 0.975]
    gam_results = {}

    # Fit the GAM model for the filtered data
    for q in quantiles:
        gam = ExpectileGAM(s(0), expectile=q)
        gam.fit(X, y)
        gam_results[f"pred_{q}"] = gam.predict(X)
    
    # Store the results in a dataframe with index that matches the original data
    predictions_gam = pd.DataFrame(gam_results).set_index(X.index)
    
    # Concatenate the results column-wise with the original data
    predictions_gam_df = pd.concat([product_data[['price', 'product', 'quantity_sold']], predictions_gam], axis=1)
    
    # Concatenate results row-wise
    all_gam_results = pd.concat([all_gam_results, predictions_gam_df], axis=0)

all_gam_results


from plotnine import ggplot, aes, geom_ribbon, geom_point, geom_line, facet_wrap, labs, scale_color_manual, theme_minimal

# Define the plot using ggplot
plot = (
    ggplot(data=all_gam_results, mapping=aes(x='price', y='quantity_sold', color='product', group='product')) +
    geom_ribbon(aes(ymax="pred_0.975", ymin="pred_0.025"), fill="#d3d3d3", color="#000000", alpha=0.75, show_legend=False) +
    geom_point(alpha=0.5) +
    geom_line(aes(y="pred_0.5"), color="blue") +
    facet_wrap('product', scales="free") +
    labs(title="GAM Price vs Quantity Model") +
    scale_color_manual(values=['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2', '#7f7f7f', '#bcbd22', '#17becf']) +
    theme_minimal() +  # Using a minimal theme for simplicity
    theme(figure_size=(10, 6))  # Adjust the figure size here (width, height)
)

# Print the plot
plot.draw()


# Optimize Price for Predicted Daily Revenue

for col in all_gam_results.columns:
    if col.startswith('pred'):
        all_gam_results['revenue_' + col] =all_gam_results['price'] * all_gam_results[col]

all_gam_results['revenue_actual'] = all_gam_results['price'] * all_gam_results['quantity_sold']

all_gam_results


best_50 = all_gam_results \
    .groupby('product') \
    .apply(lambda x: x[x['revenue_pred_0.5'] == x['revenue_pred_0.5'].max()].head(1)) \
    .reset_index(level=0, drop=True)
    
best_975 = all_gam_results \
    .groupby('product') \
    .apply(lambda x: x[x['revenue_pred_0.975'] == x['revenue_pred_0.975'].max()].head(1)) \
    .reset_index(level=0, drop=True)
    
best_025 = all_gam_results \
    .groupby('product') \
    .apply(lambda x: x[x['revenue_pred_0.025'] == x['revenue_pred_0.025'].max()].head(1)) \
    .reset_index(level=0, drop=True)


best_50


best_975


best_025


from plotnine import ggplot, aes, geom_ribbon, geom_point, geom_line, facet_wrap, labs, scale_color_manual, theme_minimal

# Define the plot using ggplot
plot = (
    ggplot(data=all_gam_results, mapping=aes(x='price', y='revenue_pred_0.5', color='product', group='product')) +
    geom_ribbon(aes(ymax='revenue_pred_0.975', ymin='revenue_pred_0.025'), fill='#d3d3d3', color='#000000', alpha=0.5, show_legend=False) +
    geom_point(aes(y='revenue_actual'), alpha=0.15, color="#2C3E50") +
    geom_line(alpha=0.5) +
    geom_point(data=best_50, color='red') +
    geom_point(data=best_975, mapping=aes(y='revenue_pred_0.975'), color='blue') +
    geom_point(data=best_025, mapping=aes(y='revenue_pred_0.025'), color='blue') +
    facet_wrap('product', scales='free') +
    labs(
        title='iPhone Case Price Optimization',
        subtitle='Maximum Median Revenue (Red) vs 95% Maximum Confidence Interval (Blue)',
        x='Price',
        y='Predicted Revenue'
    ) +
    scale_color_manual(values=['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2', '#7f7f7f', '#bcbd22', '#17becf']) +
    theme_minimal() +
    theme(figure_size=(10, 6))  # Adjust the figure size here (width, height)
)

# Print the plot
plot.draw()


best_50[['product', 'price', 'revenue_pred_0.5', 'revenue_pred_0.025', 'revenue_pred_0.975']]


# MODELING GAMS WITH EVENTS:
# --------------------------
# Essentially the same process as above, but we need to filter out the "No Promo" events
# Gets a little tricky because of limited data points for each event

# Keep the entire data as data_filtered
data_filtered = data.query('event != "No Promo"')
data_filtered


# Create a list of unique products and events
unique_products = data_filtered['product'].unique()

events_only_gam_results = pd.DataFrame()

gam_results = {}


# Loop through each product
for product in unique_products:
        
    # Filter data for the current product and event
    product_event_data = data_filtered[(data_filtered['product'] == product)]
    
    if len(product_event_data) == 0:
        continue  # skip to next iteration if no data for current product-event combination

    X = product_event_data[['price', 'event']]
    y = product_event_data['quantity_sold']
    
    # NEW: Encode the event column
    le = LabelEncoder()
    X['event'] = le.fit_transform(X['event'])
    
    # NEW: use f(1) to indicate that the event column is categorical
    gam = GAM(l(0) + f(1))
    
    gam.fit(X, y)
    gam_results[f"pred_0.5"] = gam.predict(X)
    
    # Store the results in a dataframe with index that matches the original data
    predictions_gam = pd.DataFrame(gam_results).set_index(X.index)
    
    # Concatenate the results column-wise with the original data
    predictions_gam_df = pd.concat([product_event_data[['price', 'product', 'event', 'quantity_sold']], predictions_gam], axis=1)
    
    # Concatenate results row-wise
    events_only_gam_results = pd.concat([events_only_gam_results, predictions_gam_df], axis=0)

events_only_gam_results


for col in events_only_gam_results.columns:
    if col.startswith('pred'):
        events_only_gam_results['revenue_' + col] =events_only_gam_results['price'] * events_only_gam_results[col]

events_only_gam_results

best_50 = events_only_gam_results \
    .groupby(['product', 'event']) \
    .apply(lambda x: x[x['revenue_pred_0.5'] == x['revenue_pred_0.5'].max()].head(1)) \
    .reset_index(level=0, drop=True)


best_50

	price	quantity_sold	product	event
0	12.16	1208	Standard Case \| iPhone 15 Pro Max	No Promo
1	33.87	1299	Standard Case \| iPhone 15 Pro Max	Black Friday
2	23.04	933	Standard Case \| iPhone 15 Pro Max	No Promo
3	20.45	572	Standard Case \| iPhone 15 Pro Max	New iPhone
4	17.08	1170	Standard Case \| iPhone 15 Pro Max	Christmas
...	...	...	...	...
795	60.05	127	Premium Case \| iPhone 15 Pro	New iPhone
796	68.33	175	Premium Case \| iPhone 15 Pro	No Promo
797	72.69	457	Premium Case \| iPhone 15 Pro	Black Friday
798	54.28	63	Premium Case \| iPhone 15 Pro	New iPhone
799	84.60	196	Premium Case \| iPhone 15 Pro	No Promo

	count	unique	top	freq	mean	std	min	25%	50%	75%	max
price	800.0	NaN	NaN	NaN	40.911325	20.980488	5.19	24.365	39.825	55.37	89.98
quantity_sold	800.0	NaN	NaN	NaN	632.9025	371.748974	0.0	283.0	682.5	934.75	1575.0
product	800	4	Standard Case \| iPhone 15 Pro Max	200	NaN	NaN	NaN	NaN	NaN	NaN	NaN
event	800	5	No Promo	560	NaN	NaN	NaN	NaN	NaN	NaN	NaN

	price	quantity_sold	product	event	is_event	revenue
0	12.16	1208	Standard Case \| iPhone 15 Pro Max	No Promo	0	14689.28
1	33.87	1299	Standard Case \| iPhone 15 Pro Max	Black Friday	1	43997.13
2	23.04	933	Standard Case \| iPhone 15 Pro Max	No Promo	0	21496.32
3	20.45	572	Standard Case \| iPhone 15 Pro Max	New iPhone	1	11697.40
4	17.08	1170	Standard Case \| iPhone 15 Pro Max	Christmas	1	19983.60
...	...	...	...	...	...	...
795	60.05	127	Premium Case \| iPhone 15 Pro	New iPhone	1	7626.35
796	68.33	175	Premium Case \| iPhone 15 Pro	No Promo	0	11957.75
797	72.69	457	Premium Case \| iPhone 15 Pro	Black Friday	1	33219.33
798	54.28	63	Premium Case \| iPhone 15 Pro	New iPhone	1	3419.64
799	84.60	196	Premium Case \| iPhone 15 Pro	No Promo	0	16581.60

	price	product	quantity_sold	pred_0.025	pred_0.5	pred_0.975
0	12.16	Standard Case \| iPhone 15 Pro Max	1208	936.044699	1056.475514	1154.146510
2	23.04	Standard Case \| iPhone 15 Pro Max	933	873.571104	956.955566	1052.298391
5	31.41	Standard Case \| iPhone 15 Pro Max	825	764.526030	855.147747	973.488884
6	15.92	Standard Case \| iPhone 15 Pro Max	956	891.101051	1007.301043	1130.320181
8	10.12	Standard Case \| iPhone 15 Pro Max	1180	966.403722	1076.775625	1158.523702
...	...	...	...	...	...	...
792	73.72	Premium Case \| iPhone 15 Pro	140	109.223598	179.842878	249.371987
793	63.27	Premium Case \| iPhone 15 Pro	141	142.335381	231.497810	307.703883
794	83.75	Premium Case \| iPhone 15 Pro	151	57.256986	155.666478	247.043239
796	68.33	Premium Case \| iPhone 15 Pro	175	118.528252	207.001590	273.722289
799	84.60	Premium Case \| iPhone 15 Pro	196	52.817411	153.299060	242.991816

	price	product	quantity_sold	pred_0.025	pred_0.5	pred_0.975	revenue_pred_0.025	revenue_pred_0.5	revenue_pred_0.975	revenue_actual
0	12.16	Standard Case \| iPhone 15 Pro Max	1208	936.044699	1056.475514	1154.146510	11382.303543	12846.742255	14034.421558	14689.28
2	23.04	Standard Case \| iPhone 15 Pro Max	933	873.571104	956.955566	1052.298391	20127.078235	22048.256240	24244.954919	21496.32
5	31.41	Standard Case \| iPhone 15 Pro Max	825	764.526030	855.147747	973.488884	24013.762591	26860.190736	30577.285841	25913.25
6	15.92	Standard Case \| iPhone 15 Pro Max	956	891.101051	1007.301043	1130.320181	14186.328733	16036.232608	17994.697287	15219.52
8	10.12	Standard Case \| iPhone 15 Pro Max	1180	966.403722	1076.775625	1158.523702	9780.005669	10896.969322	11724.259860	11941.60
...	...	...	...	...	...	...	...	...	...	...
792	73.72	Premium Case \| iPhone 15 Pro	140	109.223598	179.842878	249.371987	8051.963618	13258.016951	18383.702853	10320.80
793	63.27	Premium Case \| iPhone 15 Pro	141	142.335381	231.497810	307.703883	9005.559573	14646.866464	19468.424667	8921.07
794	83.75	Premium Case \| iPhone 15 Pro	151	57.256986	155.666478	247.043239	4795.272564	13037.067509	20689.871250	12646.25
796	68.33	Premium Case \| iPhone 15 Pro	175	118.528252	207.001590	273.722289	8099.035453	14144.418672	18703.444008	11957.75
799	84.60	Premium Case \| iPhone 15 Pro	196	52.817411	153.299060	242.991816	4468.352979	12969.100509	20557.107611	16581.60

	price	product	quantity_sold	pred_0.025	pred_0.5	pred_0.975	revenue_pred_0.025	revenue_pred_0.5	revenue_pred_0.975	revenue_actual
681	57.59	Premium Case \| iPhone 15 Pro	310	187.930913	281.136183	332.759844	10822.941290	16190.632764	19163.639437	17852.90
249	69.70	Premium Case \| iPhone 15 Pro Max	665	599.112739	651.307076	692.913329	41758.157914	45396.103224	48296.058997	46350.50
501	20.96	Standard Case \| iPhone 15 Pro	680	485.631098	602.293478	758.898625	10178.827809	12624.071301	15906.515173	14252.80
126	34.98	Standard Case \| iPhone 15 Pro Max	811	687.317784	783.531007	842.462477	24042.376077	27407.914622	29469.337448	28368.78

	price	product	quantity_sold	pred_0.025	pred_0.5	pred_0.975	revenue_pred_0.025	revenue_pred_0.5	revenue_pred_0.975	revenue_actual
642	41.15	Premium Case \| iPhone 15 Pro	346	219.673587	369.491219	505.314915	9039.568122	15204.563667	20793.708746	14237.90
217	55.66	Premium Case \| iPhone 15 Pro Max	766	716.977611	795.179316	875.313716	39906.973805	44259.680755	48719.961437	42635.56
509	34.58	Standard Case \| iPhone 15 Pro	300	208.717309	343.981528	473.870138	7217.444531	11894.881238	16386.429383	10374.00
162	31.90	Standard Case \| iPhone 15 Pro Max	875	755.156892	846.144311	959.223021	24089.504858	26992.003506	30599.214360	27912.50

	price	product	quantity_sold	pred_0.025	pred_0.5	pred_0.975	revenue_pred_0.025	revenue_pred_0.5	revenue_pred_0.975	revenue_actual
713	57.17	Premium Case \| iPhone 15 Pro	348	189.435146	283.062658	334.627552	10830.007280	16182.692166	19130.657121	19895.16
249	69.70	Premium Case \| iPhone 15 Pro Max	665	599.112739	651.307076	692.913329	41758.157914	45396.103224	48296.058997	46350.50
510	12.10	Standard Case \| iPhone 15 Pro	914	915.956078	1018.696826	1121.613888	11083.068541	12326.231598	13571.528040	11059.40
60	32.77	Standard Case \| iPhone 15 Pro Max	804	737.025413	828.818903	930.246918	24152.322768	27160.395455	30484.191489	26347.08

		price	product	event	quantity_sold	pred_0.5	revenue_pred_0.5
event
Black Friday	779	55.49	Premium Case \| iPhone 15 Pro	Black Friday	331	394.467150	21888.982171
Christmas	684	54.01	Premium Case \| iPhone 15 Pro	Christmas	273	399.065156	21553.509072
New iPhone	787	40.47	Premium Case \| iPhone 15 Pro	New iPhone	239	273.187263	11055.888550
The Big Game	782	47.07	Premium Case \| iPhone 15 Pro	The Big Game	255	382.429685	18000.965251
Black Friday	398	68.74	Premium Case \| iPhone 15 Pro Max	Black Friday	898	929.123711	63867.963920
Christmas	307	67.51	Premium Case \| iPhone 15 Pro Max	Christmas	724	780.239020	52673.936216
New iPhone	323	42.54	Premium Case \| iPhone 15 Pro Max	New iPhone	448	492.661971	20957.840234
The Big Game	252	65.68	Premium Case \| iPhone 15 Pro Max	The Big Game	776	752.978053	49455.598490
Black Friday	500	27.75	Standard Case \| iPhone 15 Pro	Black Friday	706	669.376975	18575.211068
Christmas	466	26.50	Standard Case \| iPhone 15 Pro	Christmas	400	635.101193	16830.181602
New iPhone	589	20.05	Standard Case \| iPhone 15 Pro	New iPhone	287	519.722904	10420.444231
The Big Game	482	24.38	Standard Case \| iPhone 15 Pro	The Big Game	572	561.684630	13693.871284
Black Friday	96	34.46	Standard Case \| iPhone 15 Pro Max	Black Friday	1347	1207.747162	41618.967204
Christmas	31	17.85	Standard Case \| iPhone 15 Pro Max	Christmas	1191	1146.311409	20461.658656
New iPhone	172	30.99	Standard Case \| iPhone 15 Pro Max	New iPhone	279	477.608573	14801.089670
The Big Game	19	32.57	Standard Case \| iPhone 15 Pro Max	The Big Game	825	932.278830	30364.321485